Yes, I concur. Not the same, but somewhat alike.
IDRC can also be software based. I found the same compaction algorithm
as IDRC in z/VSE one time. I think it was in Power Pnet, but it's been
awhile.
Tony Thigpen
Harry Wahl wrote on 7/30/22 14:06:
Tony,a
IDRC exploits the nature of IDRC compatible physical tape cartridges by writing
everything using the cartridge's internal optimal physical block size. This is
done by the cartridge drive's controller.
So, ZIP based VTS compression, such as HET, is software; while IDRC compaction
is hardware. There is no compression involved in IDRC, just hardware compaction.
IDRC uses a process called "autoblocking" to transparently optimize how much
data can fit on the cartridge's media by exploiting its optimal physical block size.
Harry
________________________________
From: IBM Mainframe Discussion List <[email protected]> on behalf of Tony
Thigpen <[email protected]>
Sent: Saturday, July 30, 2022 12:45 PM
To: [email protected] <[email protected]>
Subject: Re: AWS and IDRC/compression
I am working with my VTA vendor to reduce storage usage on the
appliance. Currently, they can compress after the unmount and uncompress
before the mount. But, this takes time, especially when servicing the
mount request if the tape is large.
I was thinking that doing an IDRC implementation, which is stream based
and performed during write/read, it might be faster even if it's not
compresses as much as with their current method.
But, if the AWS file is not compatible with IBM's implementation, then
it's going to add a step to send them the file. The current compressed
files can be uncompressed using standard linux tools.
Tony Thigpen
Jay Maynard wrote on 7/29/22 22:44:
I'm curious. What are you trying to accomplish with it? If it's just a
matter of faster transmission of entire tape images, AWS tapes compress
very well.
On Fri, Jul 29, 2022 at 8:38 PM Tony Thigpen <[email protected]> wrote:
Yes. But, it sounds like nobody else will support it as a data
interchange, so it may be unusable for us.
I will go look at it.
Tony Thigpen
Jay Maynard wrote on 7/29/22 06:38:
Are you talking about the tape data being compressed inside the AWS
image?
Hercules has a format that does this, upwardly compatible with AWS,
called
HET (Hercules Emulated Tape), but I don't know of any other
implementations
of it. Each block is compressed after being received from the program
writing the tape but before being written to the file and uncompressed
after being read but before being returned to the program reading the
tape.
On Fri, Jul 29, 2022 at 3:56 AM Tony Thigpen <[email protected]> wrote:
Does anyone know of any 'standard' for stream based (during file
creation) compression of AWS tapes?
Tony Thigpen
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
--
Jay Maynard
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
--
Jay Maynard
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
